
Comparing Two Time Periods in a CSV Without Being Fooled
Compare this month with last month from a CSV in SQL without being fooled by day counts, open periods or weekday mix, using 2 million real rows.
Everything you need to level up your coding skills.

Compare this month with last month from a CSV in SQL without being fooled by day counts, open periods or weekday mix, using 2 million real rows.

Find weekly and monthly rhythms in daily sales from a CSV: a weekday index, autocorrelation at lag 7, what a flat result looks like and how to adjust for it.

Build a cross tab of two CSV columns with DuckDB PIVOT, add row totals and percentages, and test with chi-square whether the two columns are really related.

Nine analysis moves for a CSV in SQL: cross tabs, period comparisons, concentration, seasonality, distribution shape, sampling, charts and reconciliation.

Load a CSV into MongoDB with mongoimport: what its type guessing does to zip codes, how to declare types, skip blanks and upsert on a key, with real runs.

Two ways to load a CSV into DynamoDB: the S3 import that makes a table of string attributes, and a boto3 batch writer with Decimal numbers.

Load a CSV into Oracle with SQL*Loader: a control file for quotes and dates, the bad and log files, exit codes, direct path and external tables.

Load a CSV into SQL Server with BULK INSERT: FORMAT CSV for quoted fields, FIRSTROW, UTF-8, error files, server paths and permissions.

Load CSV into ClickHouse with clickhouse-client and FORMAT CSVWithNames, tune delimiter, errors and dates, and check a file first with clickhouse-local.

Load CSV into Snowflake with a file format, a stage and COPY INTO, then read the errors: ON_ERROR, VALIDATION_MODE, VALIDATE and the 64 day skip rule.

Read and write xlsx in Node.js with exceljs: load a workbook, read typed cells, handle formulas and dates, stream large sheets and avoid common traps.

Drive a database query from cells in Excel with Power Query parameters, bind values instead of pasting text, and avoid SQL injection and stale dates.
We send one email when a new article is published. No spam.